How Would I Know If I Was Wrong?
The phrase comes from Segal's foreword, where it names the question he could not answer about his own position in the AI discourse. He believed the amplifier thesis. He still does. But he could not say what evidence would convince him he was wrong. The question is Laudanian in the deepest sense: without it, evaluation collapses into conviction, and conviction — however well-founded at the moment of its adoption — has no mechanism for self-correction. The question must be asked of every position, at every moment, by every evaluator. It is the discipline that keeps inquiry rational even when the evidence is incomplete and the stakes are high.
In the AI Story
Laudan treated the question as the operational test of intellectual seriousness. A position that cannot specify what would disconfirm it is not a rational commitment; it is a belief system. The requirement is not that disconfirming evidence exist in hand — often it does not, for reasons of underdetermination and the slow pace of empirical accumulation. The requirement is that the evaluator be able to say, in advance, what kind of evidence would force revision.
The requirement is difficult to meet. Both the triumphalist and elegist traditions, when pressed, often cannot specify what would move them. Triumphalists acknowledge burnout data but frame it as a transitional cost; they do not specify the threshold at which burnout data would compromise the amplifier thesis itself. Elegists acknowledge productivity data but frame it as capability without understanding; they do not specify the kind of evidence that would force them to accept that AI-assisted work is genuinely formative.
The inability to specify disconfirming conditions is not a neutral fact. It is diagnostic. A tradition that cannot be disconfirmed by any evidence is, in Popper's sense, not really making empirical claims. In Laudan's sense, it has retreated to the kind of ideological closure that marks degeneration. The progressive tradition is not the one that is always right; it is the one that remains open to being shown wrong, by evidence the tradition itself has specified as relevant.
Segal's honesty in the foreword — his acknowledgment that he did not know what wrong would look like, and his willingness to use this volume as an exercise in finding out — is an application of the Laudanian discipline at the scale of a book. The exercise does not guarantee he is right. It guarantees that if he is wrong, he has built the tools to discover it. That is the most that rational inquiry offers; it is also, properly understood, enough.
Key Ideas
Specify in advance. Rational commitment requires articulating what would disconfirm the position.
Disconfirming evidence need not exist yet. The requirement is specification, not immediate refutation.
Unable to specify = ideological. A position that cannot be disconfirmed is not genuinely empirical.
Self-correction requires openness. Progressive traditions retain the capacity to be shown wrong; degenerative traditions do not.
Appears in the Orange Pill Cycle
Further reading
- Karl Popper, The Logic of Scientific Discovery (1959).
- Larry Laudan, Progress and Its Problems (1977).
- Paul Meehl, "Theoretical Risks and Tabular Asterisks: Sir Karl, Sir Ronald, and the Slow Progress of Soft Psychology" (1978).